Advances in an open-source direct simulation Monte Carlo technique for hypersonic rarefied gas flows

Hypersonic vehicles that travel through rarefied gas environments are very expensive to design through experimental methods. In the last few decades major work has been carried out in developing numerical methods to capture these types of flows to a certain degree of accuracy. This accuracy is increased using particle based numerical techniques as opposed to continuum computational fluid dynamics. However, one of the modern problems of particle based techniques is the high computational cost associated with it. This thesis presents an enhanced open-source particle based technique to capture high speed rarefied gas flows. This particle based technique is called dsmcFoam and is based on the direct simulation Monte Carlo technique. As a result of the author's work dsmcFoam has become more efficient and accurate. Benchmark studies of the standard dsmcFoam solver will be presented before introducing the main advances. The results of the benchmark investigations are compared with analytical solutions, other DSMC codes and experimental data available in the literature. And excellent agreement is found when good DSMC practice has been followed. The main advances of dsmcFoam discussed are a routine for selecting collision pairs called the transient adaptive sub-cell (TASC) method and a dynamic wall temperature model (DWTM). The DWTM relates the wall temperature to the heat flux. In addition, verification and validation studies are undertaken of the DWTM. Furthermore, the widely used conventional 8 sub-cell method used to select possible collision pairs becomes very cumbersome to employ properly. This is because many mesh refinement stages are required in order to obtain accurate data. Instead of mesh refinement the TASC technique automatically employs more sub-cells, and these sub-cells are based on the number of particles in a cell. Finally, parallel efficiency tests of dsmcFoam are presented in this thesis along with a new domain decomposition technique for parallel processing. This technique splits up the computational domain based on the number of particles, such that each processor has the same number of particles to work with.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral-pg

Grantor dc:publisher.institution

University of Strathclyde

Year dc:date.issued

2013

Author dc:creator

• Ahmad, Abdul Ossman

DOI

10.48730/6bc8-4b35

Identifier

T14270

Author Identifier

200965994

OAI identifier oai:identifier

oai:strathclyde:3t945q82x